Public Support For Hosting The Olympic Summer Games In Germany: The CVM Approach - Uncg
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Archived version from NCDOCKS Institutional Repository http://libres.uncg.edu/ir/asu/
Public Support For Hosting The Olympic Summer Games In
Germany: The CVM Approach
By: Pamela Wicker, John C. Whitehead, Daniel S. Mason, & Bruce K. Johnson
Abstract
Cities and their respective regions must weigh the merits of hosting major sport events, including the Olympic Games.
This paper presents a contingent valuation method estimate of the monetary value of intangible benefits to Germans of
hosting the Olympic Summer Games. In a nation-wide online survey, 6977 respondents said whether they would
support a referendum to host the Games. The survey employed a payment card format containing monthly tax amounts
to elicit individual willingness-to-pay for the Games over a five-year period. In the weighted sample, 26% expressed an
average willingness-to-pay of e 51. Willingness-to-pay varied widely across regions. Around Cologne, the average
willingness-to-pay was e 100. Interval data hurdle models reveal that policy consequentiality and various intangible
benefits increased willingness-to-pay. Aggregate willingness-to-pay over a five-year period amounted to e 46 billion
which exceeded the estimated costs of the 2024 Summer Games for Hamburg. The findings have implications for policy
makers since they show what regions within Germany most support hosting the Games.
Wicker, P., et al. (2016). "Public support for hosting the Olympic Summer Games in Germany: The CVM approach."
Urban Studies 54(15): 3597-3614. https://doi.org/10.1177/0042098016675085. Publisher version of record available at:
http://journals.sagepub.com/doi/full/10.1177/0042098016675085Public support for hosting the
Olympic Summer Games in
Germany: The CVM approach
Pamela Wicker
German Sport University Cologne, Germany
John C Whitehead
Appalachian State University, USA
Daniel S Mason
University of Alberta, Canada
Bruce K Johnson
Centre College, USA
Abstract
Cities and their respective regions must weigh the merits of hosting major sport events, including
the Olympic Games. This paper presents a contingent valuation method estimate of the mone-
tary value of intangible benefits to Germans of hosting the Olympic Summer Games. In a nation-
wide online survey, 6977 respondents said whether they would support a referendum to host the
Games. The survey employed a payment card format containing monthly tax amounts to elicit
individual willingness-to-pay for the Games over a five-year period. In the weighted sample, 26%
expressed an average willingness-to-pay of e 51. Willingness-to-pay varied widely across regions.
Around Cologne, the average willingness-to-pay was e 100. Interval data hurdle models reveal
that policy consequentiality and various intangible benefits increased willingness-to-pay. Aggregate
willingness-to-pay over a five-year period amounted to e 46 billion which exceeded the estimated
costs of the 2024 Summer Games for Hamburg. The findings have implications for policy makers
since they show what regions within Germany most support hosting the Games.
Keywords
contingent valuation method, host city, intangible benefit, Olympic Games, willingness-to-pay
Corresponding author:
Pamela Wicker, German Sport University Cologne,
Department of Sport Economics and Sport Management,
Am Sportpark Muengersdorf 6, Cologne, NRW, 50933,
Germany.
Email: p.wicker@dshs-koeln.deIntroduction ‘Public support is crucial in supporting
Despite substantial costs and questionable cultural activities and, more generally, for
economic benefits, cities willingly compete to developing the civic amenities that become
host major sporting events (Burbank et al., ‘‘key’’ assets of the city’ (Begg, 1999: 800).
2001). While this may reflect pro-growth city Presumably, local support for hosting the
elites pursuing their interests (Amin, 2003; Olympics should depend on the benefits,
Thornley, 2002), some research shows that both tangible and intangible, voters expect
cities may receive intangible, non-pecuniary to receive. Research on tangible benefits of
benefits from hosting sport teams and events hosting, such as economic impact or labor
(Santo, 2008) and that ‘the Olympic Games market outcomes of the Games (Baade et al.,
represent the biggest prize for cities seeking 2010; Feddersen and Maennig, 2013) has
mega-events’ (Holcomb, 1999: 59). established that benefits are often overstated,
In March 2015, the German Olympic particularly when estimated ex ante (Porter
Sports Confederation decided to submit a and Fletcher, 2008). Consequently, tangible
bid for Hamburg to host the 2024 Olympic economic benefits cannot justify local sup-
Summer Games. Because the International port for event hosting.
Olympic Committee (IOC) makes prospec- However, tangible benefits are only part
tive cities consider public opinion before of the story. Intangible benefits of sport
they bid, Hamburg held a referendum on events include national and civic pride, gen-
hosting the Games in late November 2015. eral feel-good-factors and image improve-
The outcome of the referendum was a ‘no’ ments (Johnson, 2008; Kavetsos and
vote. Elsewhere in Germany, a lack of public Szymanski, 2010; Rowe and McGuirk,
support stalled Munich’s chance to host the 1999). These intangibles are public goods
2018 and 2022 Winter Olympic Games, most characterised by non-excludability and non-
recently in 2013 when citizens of Munich rivalry (Downward et al., 2009). Since quan-
and three nearby districts failed to support a tities and prices cannot be observed for
2022 bid. The question arises, to what extent non-traded public goods, it is difficult to
do Germans favor hosting Olympic Games, assign monetary values to them. The contin-
and which cities/regions support it? gent valuation method (CVM) can do so(Carson, 2000). A CVM survey presents traditional economic perspective has been
respondents with a hypothetical scenario that consumers must directly use a good to
and asks to state their willingness-to-pay get utility from it – its monetary value is its
(WTP) for the public good described in the use value. But consumers may also get utility
scenario (Carson, 2000). without physically using a good. That utility
The purpose of this study is to estimate is referred to as passive-use or non-use value
the monetary value of intangible benefits to (Carson, 2000). The intangible benefits of
Germans of hosting the Olympic Summer hosting sport events noted earlier fall into
Games using CVM. We advance three main this category. People need not purchase tick-
research questions: (1) what are the individ- ets and attend Olympic events to get utility
ual and aggregate WTP of the German pop- from them (Humphreys et al., 2016).
ulation for hosting the Summer Olympics? For example, Kavetsos and Szymanski
(2) Which German regions value hosting the (2010) showed that hosting football events
Games the most? And (3) what factors influ- creates a feel-good factor. Pawlowski et al.
ence individual WTP? Data were collected (2014) concluded that hosting events raises
nationwide from December 2013 to March subjective wellbeing even more than the
2014 using an online survey. It asked respon- pride from sporting success. Hiller and
dents to state their likelihood of supporting Wanner (2014) documented that Olympic
a referendum to host the Summer Games hosting produces psycho-social benefits and
and presented them with a payment card that the festival atmosphere positively affects
containing monthly amounts of a five-year the public mood. This study estimates the
tax to finance the Games. Our findings have monetary value of these intangible benefits
policy implications since they show which from hosting Olympic Games.
regions support hosting the Olympics. This
study contributes to the developing body of
stated preference literature within urban
Incentive compatibility and
studies. consequentiality in CVM
The following section discusses CVM, Carson and Groves (2007) and Carson
passive-use values, incentive compatibility (2012) argued that stated preference surveys
and consequentiality, along with previous with incentive compatible, consequential
CVM studies on hosting the Olympics, valuation questions produce accurate WTP
and outlines this study’s contribution. The estimates. A question is incentive compatible
methods section describes the questionnaire, if it induces respondents to reveal their true
the sample and empirical analysis strategy. preferences, i.e., respondents answer in the
After presenting and discussing the results, same way as if they were making a real
the article finishes with some concluding choice or payment (Carson et al., 2014;
remarks. Cummings et al., 1995). With incentive com-
patible questions, respondents neither
underestimate nor overestimate their WTP,
Conceptual framework and avoiding hypothetical bias. For example, a
previous research findings referendum with majority rule is incentive
compatible because there are no incentives
CVM and passive-use values for strategic response. In contrast, an open-
The CVM is ‘the only feasible method for ended question, e.g. how much are you will-
including passive-use considerations in an ing to pay?, invites strategic behaviour.
economic analysis’ (Carson, 2000: 1413). A Respondents who want the policy, but donot think they will actually pay may exagge- value (Carson et al., 2014; Landry and List,
rate their WTP in order to increase the prob- 2007; Vossler and Evans, 2009). Vossler and
ability that the policy will pass. Those who Watson (2013) and Groothuis et al. (2015)
desire the policy, but do not want to pay, found that perceived inconsequentiality
may understate their true WTP so they can reduces stated WTP.
free ride. Those willing to pay an amount
greater (less) than the cost of the policy Application of CVM in sport in an urban
should vote in favour of (against) the policy. context
Incentive compatibility is also affected by
the payment vehicle (Wiser, 2007): a collec- The CVM has been used to estimate the mone-
tive and coercive payment mechanism such tary value of public goods in an urban context.
as a tax increase is considered more incen- For example, Johnson and Whitehead (2000)
tive compatible than a voluntary payment applied CVM to stadiums and Johnson et al.
mechanism, e.g. donation to a voluntary (2001) to professional sports teams. Other
studies followed, with Barlow and Forrest
fund, because expectations about the pay-
(2015) estimating WTP for small-town football
ments of others affect WTP (Wiser, 2007).
clubs and Harter (2015) for a town arena.
While a tax ensures that all citizens pay, a
Johnson et al. (2012) examined WTP for a new
donation invites free riding because people
arena with non-sport amenities to enhance
may benefit from the potential supply of the
downtown vibrancy.
good without paying. Moreover, the type
and context of a WTP question is critical:
for example, dichotomous choice questions CVM studies on hosting the Olympic
are not considered incentive compatible Games
(Cummings et al., 1995), while referendum This section reviews existing studies estimat-
questions have been found to be incentive ing the monetary value of public goods cre-
compatible (Carson et al., 2014). ated by hosting the Olympic Games. Two
A question is consequential if respondents studies focused on London’s bid for the 2012
believe their response might affect something Summer Olympics; London’s bid was sup-
they care about (Groothuis et al., 2015). It is ported by the British government following a
assumed that ‘hypothetical behavior will be decision in April 2003. In September 2004,
similar to real behavior if there is a positive Atkinson et al. (2008) conducted face-to-face
chance that the hypothetical behavior will interviews in London (n = 602), Manchester
have real consequences’ (Groothuis et al., (n = 151) and Glasgow (n = 152). The
2015: 4). A perceived lack of consequential- authors acknowledged that the Manchester
ity can lead to protest responses (Groothuis and Glasgow samples were not representative,
and Whitehead, 2009). but did not comment on whether the London
Carson and Groves (2007) argued that sample was representative or whether any
respondents have little incentive to invest corrective measures were taken. Their CVM
effort in an inconsequential survey, but do payment vehicle was an increase in the house-
have incentives in a consequential survey to hold’s annual council tax bill. They elicited
respond truthfully. While theory makes no WTP using a payment card. Average house-
predictions about the effect of consequenti- hold WTP in London (£22) exceeded those in
ality on WTP, mounting empirical evidence Manchester (£12) and Glasgow (£11).
from laboratory and field experiments indi- The decision to award the 2012 Olympics
cates that consequential hypothetical ques- to London was made by the IOC on 6 July
tions lead to more accurate statements of 2005. Walton et al. (2008) conducted face-to-face surveys in public areas in Bath (n = 167) seven answer categories elicited WTP. With
from end of April to beginning of July 2005. 63% males and an average age of 19.6 years,
Comparison with census data shows that the the final sample (n = 1768) is not represen-
sample is representative of Bath’s population. tative of the population, but no corrective
But the survey timing, just weeks before the actions were taken. Average individual WTP
decision to give the Games to London, sug- of the full sample amounted to US$138,
gests that perceived policy consequentiality ranging from US$93 in the Mountain region
was probably low. The payment vehicle was a to US$151 in the Mid Atlantic.
national fixed-rate tax. WTP was elicited in a Preuss and Werkmann (2011) estimated
double-bounded dichotomous choice format. the experiential value of hosting the 2018
Mean individual WTP was £70 and aggregated Winter Olympic Games in Munich, Germany.
WTP was approximately £5.8 million for Bath, They interviewed 1011 people in public places
supporting the notion that non-Londoners val- in the Rhine–Main-region, which is at least
ued public goods from London hosting the three hours away from Munich. Males and
Olympics. young people were overrepresented, but no
In 2007, Heisey (2009) conducted public corrective measures were taken. The CVM
intercept surveys to determine WTP for scenario did not specify a payment vehicle. A
hosting the 2016 Summer Olympics in Berlin payment card was used to elicit individual
(n = 499), San Francisco (n = 544), and WTP. Mean WTP was not reported.
Chicago (n = 342). Chicago was added after Aggregate WTP was between e 617 million to
San Francisco dropped its bid during the e 803 million.
data collection phase. The CVM payment These previous studies also examined the
vehicle was a donation to a private organisa- determinants of WTP. Males reported higher
tion. A payment card was used to elicit WTP than females (Coates and Szymanski,
WTP. The sample structure was not com- 2015; Walton et al., 2008). Age had a negative
pared with the cities’ population, although effect in one study (Walton et al., 2008), and
the average age of the respondents, i.e., was insignificant in others (Atkinson et al.,
between 32 and 37 years, seems relatively 2008; Coates and Szymanski, 2015; Preuss
young. Average individual WTP in Chicago and Werkmann, 2011). Income had a positive
(US$55) exceeded those in San Francisco effect on WTP (Atkinson et al., 2008; Coates
(US$36) and Berlin (e 16). Aggregate and Szymanski, 2015; Heisey, 2009; Preuss
yearly WTP was e 81.75 million in Berlin, and Werkmann, 2011; Walton et al., 2008), as
US$154.15 million in San Francisco, and did a general interest in sport and organised
US$439.1 million in Chicago. exercise (Preuss and Werkmann, 2011;
In April 2014, Coates and Szymanski Walton et al., 2008). People expecting benefits
(2015) examined the WTP to host the 2024 from the event (Preuss and Werkmann, 2011)
Summer Olympics in the USA, when four had a higher WTP, even more when they
cities were trying to be the US bidder. North thought intangible benefits would be more
American University students organised the important than tangible benefits (Atkinson
data collection. The sample consisted of peo- et al., 2008).
ple paid ten cents to complete the survey
through a Mechanical Turk employment
Shortcomings of previous research and
site. The CVM scenario proposed that the
Games would be hosted in the respondent’s contribution of this study
region. The payment vehicle was a one-time Several shortcomings can be observed in the
lump sum tax. A closed-ended question with literature. First, some of the studies made noattempt to compensate for unrepresentative questionnaire could only be completed once
samples. Second, some of the samples were per internet protocol. The survey attracted
on the small side of the ‘several hundred to a 7721 respondents. During the data cleaning
couple thousand observations [.] generally process, people younger than 16 were
required to achieve reasonable reliability removed as well as those clicking through
from a sampling (confidence interval) per- the survey without thinking, e.g., completing
spective’ (Carson, 2000: 1416). Third, policy a six-page questionnaire in less than two
consequentiality has not yet been considered. minutes, choosing the same answer to sev-
The present study addresses these shortcom- eral consecutive questions, those providing
ings with a large, weighted sample and by implausible, e.g. invalid postcodes, educa-
using incentive-compatible, consequential tional level implausible for a certain age, or
survey questions. Another contribution is nonsensical answers, e.g., tenure greater
that we examine the effect of consequential- than age, and those with theoretically invalid
ity by decomposing WTP into extensive WTP responses such as increasing likelihood
(whether WTP is positive) and intensive (the of WTP with increasing payments
magnitude of WTP) margins. (Dickinson and Whitehead, 2015). The final
sample contained 6977 observations.
Method
Questionnaire and variables
Data collection Table 1 provides variable names and defini-
Primary data were collected in Germany tions. Since ordering effects may affect WTP
with a nationwide online survey from in CVM studies including multiple valuation
December 2013, a month after the Munich scenarios (Johnson et al., 2006), respondents
referendum to host the 2022 Winter were randomly assigned either the two foot-
Olympics failed, to March 2014. In the after- ball scenarios (which were presented conse-
math of Munich, German officials shifted cutively) or the Olympic scenario first
their efforts to attract the Summer Games, (FIRST). The Olympic section started with
without specifying a host city. Nevertheless, questions about respondents’ general inter-
the failure of the referendum so soon before est in sport. They were asked to state their
our survey may have affected responses. level of agreement with a set of statements
This study is part of a larger project on a five-point Likert scale, from strongly
examining the value of sport, specifically disagree (1) to strongly agree (5). The state-
football (soccer; Wicker et al., 2016) and ments referred to the respondents’ interest in
Olympic Games, to the German population. sport in general (INTEREST); regular sport
Sosci Survey (www.soscisurvey.de) hosted practice (PLAY); whether they identify with
the survey. The survey link was distributed Germany (IDENTIFY); whether they feel
using social media, e.g., Facebook, Twitter, proud (PROUD) and happy (HAPPY)
city websites and local radio websites when German athletes and teams succeed at
throughout Germany. Because of the foot- international competitions; whether they
ball questions, the link was also posted in think Germany’s reputation is burnished by
fan forums and fan club websites. This sam- German success at international competi-
pling procedure resulted in a convenience tions (REPUTATION); and whether they
sample rather than a representative random regard German athletes as role models
sample. (ROLEMODELS). These variables were
Survey participation was limited to coded as 1 if respondents replied strongly
respondents aged 16 years or older, and the agree or agree and 0 otherwise.Table 1. Overview of variables.
Variable Description
PRESTIGE Do you think Germany’s prestige and standing in the world would rise if Germany
hosted the Summer Olympics? (1 = yes)
POS_WTP Positive WTP (1 if WTP . e 10)
WTP Amount of willingness-to-pay (in e ) if POS_WTP = 1
FIRST Order of Olympics scenario (1 = first scenario; 0 = third scenario)
INTEREST I am interested in sports in general (strongly agree/agree = 1)
PLAY I practice sport regularly, i.e., at least once per week (strongly agree/agree = 1)
IDENTIFY I identify with Germany (strongly agree/agree = 1)
PROUD I am proud when German athletes/teams are successful at international sport
competitions (strongly agree/agree = 1)
HAPPY I am happy when German athletes/teams are successful at international sporting
competitions (strongly agree/agree = 1)
REPUTATION It is important for the reputation of Germany that German athletes/teams are
successful at international sporting competitions (strongly agree/agree = 1)
ROLEMODELS German athletes are role models (strongly agree/agree = 1)
REFERENDUM A referendum is a good way for citizens to express their preferences for sport in
Germany (strongly agree/agree = 1)
SHARED I believe the results of this survey will be shared with policy makers (strongly agree/
agree = 1)
AFFECT I believe the results of this survey could affect decisions on sport in Germany
(strongly agree/agree = 1)
INFO I understand all of the information presented to me in this survey (strongly agree/
agree = 1)
CONFIDENCE I have confidence in the ability of the German government to achieve the goals of
sport policy (strongly agree/agree = 1)
INCOME Individual monthly net income (from 1 = up to e 500 to 9 = over e 4000)
MALE Gender of the respondent (1 = male, 0 = female)
AGE Age of the respondent (in years)
TENURE Number of years living in the city
SCHOOLING Years of schooling
POSTCODE Postcode area within Germany (from 0 to 9)
Afterwards, the Olympic scenario was Suppose the German Olympic Sports
presented: Confederation considers submitting a bid
for Olympic Summer Games in the future.
The Olympic Games may be the most famous To pay the extra cost of hosting the
and prestigious of all sporting competitions in Olympics, the German government would
the world. Some cities have hosted multiple impose a monthly income tax surcharge on
Olympic Games. For instance, in 2012, each individual for the next 5 years.
London hosted its third Summer Olympics.
Up to now Munich is the only German city
Given the possibility of temporal embedding
which has hosted Olympic Games after World
War II, i.e., the 1972 Summer Olympics. Its
effects in CVM surveys (Johnson et al.,
bid to host the 2018 Winter Olympics was 2006), the duration of the monthly tax pay-
unsuccessful. Moreover, the outcome of the ment, five years, was clearly specified. The
referendum in the Munich region in November survey opened a few weeks after the Munich
2013 was that people did not support another referendum failed and before any bid cities
bid of Munich for Winter Olympics. had been mentioned in the media. Because itwas not clear if one of the German cities midpoint of a very likely vote and next high-
which applied previously for Olympic est amount at some less likely response. For
Summer Games, i.e., Berlin for the 2000 example, if a respondent stated very likely to
Games, Leipzig for the 2012 Games, or a city e 10 and somewhat likely to e 25, then WTP
that was eliminated in an intra-German com- was coded at e 17.50.
petition in 2003, i.e., Stuttgart, Frankfurt, After the scenario, respondents stated
Düsseldorf, Hamburg, or another city would their level of agreement with several policy
be the preferred German bidder, the survey related statements on a five-point Likert
did not specify a host city. This avoided scale. They were asked whether they thought
the possibility that the German Olympic a referendum would be a good way for
Sports Confederation might name different citizens to express their preferences
potential hosts during the survey period, (REFERENDUM); whether they believed
adversely affecting responses. Since the sur- the survey results would be shared with pol-
vey was also administered to citizens of the icy makers (SHARED); whether they
Munich region, assuming that the Olympics believed the results would affect decisions on
would be hosted close to the respondents’ sport in Germany (AFFECT); and whether
home would not be realistic given the failed they had confidence in the ability of the
referendum. German government to achieve the goals of
Respondents were then asked if they sport policy (CONFIDENCE). Moreover, a
thought Germany’s world prestige and statement capturing whether respondents
standing would rise if Germany hosted the understood the information in the survey was
Summer Olympics (PRESTIGE). Then they included (INFO). These variables were
were asked, ‘Suppose that this proposal was recoded as 1 when the respondents replied
put to a referendum vote for all Germans . strongly agree or agree and 0 otherwise. The
how likely do you think it is that you would questionnaire finished with questions about
vote for the proposal at the following socio-economic characteristics including
tax amounts?’. Respondent’s likelihood of gender (MALE), age (AGE), education
voting for specific tax amounts was assessed (SCHOOLING), postcode (POSTCODE),
on a five-point scale, from very unlikely (1) number of years lived in the current city
to very likely (5). Respondents were pre- (TENURE), and personal monthly net income
sented with a payment card including seven (INCOME). For income, the midpoint Euro
different tax amounts (e 10, e 25, e 50, e 100, value of the respective income category was
e 150, e 200 and e 250). This question used (Liebe, 2007). The natural logarithm was
eliciting individual WTP was incentive- deemed appropriate because its distribution is
compatible and is expected to result in more closer to the normal distribution.
conservative WTP estimates compared with
a dichotomous choice referendum question.
Two variables resulted from the scenario. Initial sample and use of weights
The first is POS_WTP which is equal to 1 if In the initial sample, 76% of the respon-
the respondent is very likely to support the dents were male and the average age was
referendum at e 10 or higher. Respondents 31.7 years. In comparison, 48% of the
who said they would only be somewhat likely German population aged 15 or older is
to vote in favour were coded as no votes to male, and the average age is 44 years
mitigate hypothetical bias (Loomis, 2011). In (Federal Statistical Office, 2015). Thus,
the case of POS_WTP = 1, the second vari- males and younger people are over-
able is WTP which was coded at the represented in the initial sample.Several circumstances may explain the the proportion of males. The mean values of
non-representative sample. First, the link to other variables changed little.
the survey informed potential respondents The problem of non-responses could not
that the survey was about sports, a topic of be fully addressed. The sampling strategy
greater interest to young males (Downward and the anonymity of the survey compro-
and Rasciute, 2010). Bundling Olympic and mise the use of measures to adjust for non-
football topics may have intensified the self- responses. One common option for dealing
selection of younger males, since the typical with non-response is to use sample weights
German football fan is a young male based on different selection probabilities
(Schmidt and Högele, 2011). Second, the (Carkin and Tracy, 2014). But this procedure
sampling strategy and use of an online sur- could not be applied. Because our sample
vey may have affected the sample structure. was anonymous, we do not know the charac-
We chose an online survey over mail surveys teristics of non-respondents. Anonymous
for several reasons: online surveys are faster sampling also prevents us from reminding
and cheaper (Dillman et al., 2014); online non-respondents to respond, a common pro-
surveys can require answers before allowing cedure when identities of survey recipients are
respondents to move to the next question, known (Hansen et al., 2014). While these pro-
reducing the likelihood of incomplete cedures could not be applied, weighting the
responses; and the survey design required sample partially accounts for non-response
randomising the order of scenarios, which is and drop-outs and, thus, mitigates sample bias
easy to do in online surveys. (Höfler et al., 2005). Altogether, this large data
Disadvantages of online convenience sam- set allows an interesting and valuable analysis
pling include missing people who are less in terms of originality, significance, and rigour
inclined or able to participate in online sur- of the underlying research.
veys, which might have contributed to such a
young and male sample. Although a large Empirical analysis
part of the population is online nowadays, the The empirical analysis is based on the
extent of Internet usage still differs between weighted sample. Two empirical issues must
population groups. Recent research shows be considered. First, given its prominence in
that younger people and males are more the CVM scenario, PRESTIGE is a poten-
active online than older people and females tially endogenous variable. We attempted
(Drabowicz, 2014; van Deursen et al., 2015) several models for estimating the effect of
which may explain the difference between the PRESTIGE on WTP including an instru-
initial sample and the German population. mental variable model (Pawlowski et al.,
To compensate, we weighted the sample 2014), but could not identify suitable instru-
for age, gender, and region (postcode area). ments when using the sample with regional
Since sample size and composition are criti- weights. As an alternative we estimated a
cal to the reliability of CVM studies, using bivariate probit model with PRESTIGE and
appropriate weights is recommended for POS_WTP as the dependent variables – sim-
non-representative samples (Carson, 2000). ilar to Morgan and Whitehead (2015). The
We calculated weights based on detailed pop- bivariate probit model estimates the correla-
ulation statistics provided by the German tion in error terms which captures the corre-
Federal Statistical Office (2015). Table 2 lation in unexplained variation.
reports characteristics of the unweighted and Second, some applications of the CVM
weighted sample. Using weights increased lead to a large number of zero WTP
average age, income and tenure, and reduced responses. A Tobit model could beTable 2. Summary statistics (n = 6977).
Unweighted sample Weighted sample
Variable Mean SD Mean SD Min Max
PRESTIGE 0.66 – 0.59 – 0 1
POS_WTP 0.28 – 0.26 – 0 1
WTP 49.71 62.04 51.38 94.30 17.5 275
FIRST 0.50 – 0.51 – 0 1
INTEREST 0.86 – 0.79 – 0 1
PLAY 0.64 – 0.58 – 0 1
IDENTIFY 0.75 – 0.71 – 0 1
PROUD 0.70 – 0.67 – 0 1
HAPPY 0.68 – 0.64 – 0 1
REPUTATION 0.65 – 0.63 – 0 1
ROLEMODELS 0.72 – 0.69 – 0 1
REFERENDUM 0.53 – 0.51 – 0 1
SHARED 0.49 – 0.53 – 0 1
AFFECT 0.28 – 0.28 – 0 1
INFO 0.81 – 0.80 – 0 1
CONFIDENCE 0.26 – 0.23 – 0 1
INCOME 1647.74 1213.71 1970.90 1233.15 250 4250
MALE 0.76 – 0.48 – 0 1
AGE 31.70 12.30 44.53 15.51 16 99
TENURE 19.38 14.38 26.47 18.43 0 90
SCHOOLING 13.74 2.91 13.55 3.23 5 17
POSTCODE
0 0.02 – 0.08 – 0 1
1 0.03 – 0.09 – 0 1
2 0.04 – 0.10 – 0 1
3 0.04 – 0.11 – 0 1
4 0.22 – 0.12 – 0 1
5 0.11 – 0.11 – 0 1
6 0.15 – 0.09 – 0 1
7 0.11 – 0.11 – 0 1
8 0.18 – 0.10 – 0 1
9 0.10 – 0.09 – 0 1
estimated, but whether a respondent would (del Saz-Salazar and Rausell-Köster, 2008).
be willing to pay anything at all may depend Thus, a hurdle model was preferred.
upon different factors than does the amount With our data, where the continuous
of payment (Castellanos et al., 2011). If so, a WTP is estimated between intervals, the
hurdle model should be estimated, where a grouped or interval data hurdle model is
probit model is fit to the first decision and appropriate (Cameron and Huppert, 1989).
a continuous regression model is fit to the A comparison with an OLS model supported
second. Hurdle models are useful when the choice of the interval regression. Our
extensive and intensive margins are investi- final model can be described as follows:
gated using data with a large proportion of Bivariate Probit : PRESTIGE = fðXÞ;
zero responses (Castellanos et al., 2011;
POS WTP = fðXÞ, r ð1Þ
Humphreys et al., 2010). In this case Tobit
models would lead to misleading inferences Interval regression : LNðIWTPÞ = fðXÞ ð2Þwhere r is the correlation in error terms examines the determinants of prestige and a
between PRESTIGE and POS_WTP, IWTP positive WTP, while the interval regression
= 1 if WTP \ 25, 2 if 25 WTP \ 50, 3 if estimates the drivers of the amount of stated
50 WTP \ 100, 4 if 100 WTP \ 150, 5 WTP. The bivariate probit model shows that
if 150 WTP \ 200, 6 if 200 WTP \ PRESTIGE and POS_WTP are both
250, 7 if WTP 250. Hence, IWTP was affected by various intangible benefits such
obtained by transferring the WTP Euro as pride derived from sporting success,
values calculated earlier into numbers from enhanced reputation from hosting the
1 to 7. Games, and viewing athletes as role models.
Previous research has found that respon- Anticipation of positive and intangible
dents who think that the survey results will effects was also positively associated with
affect policy decisions, i.e., the survey is con- WTP in previous research (Atkinson et al.,
sequential, are willing to pay significantly 2008; Preuss and Werkmann, 2011).
more than others (Groothuis et al., 2015). Moreover, playing sports, identification
Our measurement of consequentiality and with Germany, and confidence the German
decomposition of WTP with the hurdle government can achieve its sport policy
model provides additional insights into this goals significantly increase the probability of
result. We included measures for whether a positive WTP, while simultaneously reduc-
respondents think the survey results will be ing the WTP amount. This change in the
shared with decision makers (SHARED) sign on the coefficient can also be observed
and whether the results will affect decisions for the reputation and confidence variables.
(AFFECT). Previous research also documented that pos-
itive WTP and amount of WTP are affected
by different factors (Liebe et al., 2011), sup-
Results and discussion porting the two-fold decision. Our models
The summary statistics (Table 2) of the reveal that Germans support hosting the
weighted sample show that 26% would be Olympics, but object to funding it through
willing to pay a monthly income tax over a taxes, suggesting that the payment vehicle
five-year period of at least e 10 to host the plays a role. Recent examples of inefficient
Olympic Summer Games. For those willing government spending, e.g., Airport Berlin-
to pay, average individual WTP is e 51, Brandenburg, Hamburg Elb Philharmonics,
much higher than the e 16 found by Heisey Stuttgart 21, have led to a negative attitude
(2009), who estimated WTP for Berlin host- in general towards large-scale projects in
ing the 2016 Summer Games. However, his Germany (Könecke et al., 2016). Hence,
sample was smaller, not representative, and many Germans may oppose paying more
not weighted. Our result is similar to the taxes to the government for such projects.
average e 46 WTP identified by Wicker et al. Breuer and Hallmann (2011), for instance,
(2015) for Olympic medal success, i.e., for found that only 15.3% of respondents pre-
Germany being ranked first in the final ferred the government to support Olympic
medal table. Thus, the intangible benefits athletes, while 43.4% favoured a sport feder-
from hosting Olympic Games are of similar ation and 33.5% preferred a foundation.
magnitude as the intangible benefits from Not only trust in the institution taking
medal success. The statistics of the remain- care of the money may play a role, but also
ing variables can be obtained from Table 2. the universality of a payment vehicle.
The regression results are summarised in Income taxes, used here and in prior studies,
Table 3. The bivariate probit model are relevant to people participating in theTable 3. Determinants of prestige and WTP (weighted sample; bivariate probit and interval regression
model).
Bivariate probit Interval regression
PRESTIGE POS_WTP LN(WTP)
Coeff. t Coeff. t Coeff. t
Intercept 20.554* 23.45 21.73* 29.86 1.609* 4.34
FIRST 20.106* 23.26 0.309* 8.81 20.474* 26.86
INTEREST 20.059 21.33 20.055 21.10 20.049 20.48
PLAY 20.090* 22.54 0.075* 1.96 20.261* 23.47
IDENTIFY 0.229* 5.06 0.270* 5.25 20.222* 21.98
PROUD 0.133* 2.55 0.173* 3.15 0.058 0.49
HAPPY 20.066 21.39 0.147* 2.92 0.236* 2.21
REPUTATION 0.520* 13.31 0.150* 3.47 20.300* 23.37
ROLEMODELS 0.346* 9.12 0.232* 5.45 0.352* 3.73
REFERENDUM 0.108* 3.21 0.138* 3.80 20.060 20.84
SHARED 20.033 20.90 20.008 20.20 0.019 0.24
AFFECT 0.022 0.55 0.104* 2.53 0.200* 2.54
INFO 0.014 0.33 0.122* 2.65 0.211* 2.20
CONFIDENCE 0.291* 6.98 0.219* 5.26 20.158* 22.04
LN(INCOME) 0.013* 3.58 0.023 0.98 0.189* 4.14
MALE 0.112* 3.13 20.057 21.46 0.274* 3.55
AGE 20.010* 27.55 20.003* 22.29 20.013* 24.51
TENURE 0.0004 20.37 0.001 0.92 0.005* 2.18
SCHOOLING 20.003 20.47 20.009 21.58 0.040* 3.48
POSTCODE_0 20.108 21.04 20.102 20.88 20.014 0.28
POSTCODE_1 20.072 20.82 0.240* 2.57 0.283 1.56
POSTCODE_3 0.185* 2.25 20.266* 22.93 20.554* 22.74
POSTCODE_4 0.038 0.50 0.265* 3.13 20.167 20.97
POSTCODE_5 20.069 20.87 0.093 1.08 0.129 0.74
POSTCODE_6 20.169* 22.18 0.148 1.73 20.217 21.26
POSTCODE_7 0.049 0.62 0.037 0.42 20.237 21.33
POSTCODE_8 20.130 21.61 0.117 1.31 0.318 1.78
POSTCODE_9 20.005 0.06 0.127 1.35 0.278 1.51
Rho 0.494* 25.03
Sigma 1.252* 34.26
LL 27618.36 22439.00
n 6977 6977 1997
Note: *p \ 0.05; reference is POSTCODE_2 (Hamburg region).
labour force and to retirees, as pensions are German sport policy are significantly more
also taxed in Germany. Nevertheless, the sur- likely to report a positive WTP.
vey may have been perceived as less conse- Furthermore, WTP is 22% higher when
quential by people not drawing a salary or a respondents think it will affect policy. The
pension, or whose low incomes are not taxed. positive relationship between policy conse-
While we do not control for employment sta- quentiality and WTP is consistent with pre-
tus in our models, the income effect should be vious research (Groothuis et al., 2015;
able to explain some of this variation. Vossler and Watson, 2013); yet these studies
Regarding consequentiality, respondents used different measures of consequentiality.
who think this survey’s results will affect The insignificance of the SHARED variableTable 4. WTP by postcode area and aggregate WTP.
Postcode area POS_WTP WTP Total Aggregate
population WTPa
(included states Mean Mean SD (in million) (in million e )
and cities)
0 (Saxony incl. Leipzig, Eastern 0.171 38.91 38.79 6.4 1917
Thuringia, Southern Saxony-Anhalt,
Southern Brandenburg)
1 (Berlin, Brandenburg, Mecklenburg- 0.285 71.33 125.64 7.0 4963
West Pomerania)
2 (Hamburg, Bremen, Schleswig- 0.203 51.31 94.31 8.6 3974
Holstein, Northern Lower Saxony)
3 (Southern Lower Saxony incl. 0.198 31.11 59.90 8.8 2398
Hanover, Eastern North Rhine-
Westphalia, Northern Hesse,
Northern Saxony-Anhalt)
4 (Northwestern North Rhine- 0.316 38.85 68.51 10.1 5543
Westphalia incl. Düsseldorf, Western
Lower Saxony)
5 (Southern North Rhine-Westphalia 0.267 100.01 154.65 9.1 10,847
incl. Cologne; Northern Rhineland
Palatinate)
6 (Southern Hesse incl. Frankfurt, 0.299 34.94 57.65 7.5 3512
Southern Rhineland Palatinate,
Saarland)
7 (Baden-Württemberg incl. Stuttgart) 0.252 35.60 60.62 8.7 3495
8 (Southern Bavaria incl. Munich) 0.276 55.34 94.62 7.9 5393
9 (Northern Bavaria incl. Nuremberg; 0.245 56.47 102.91 7.0 4367
Western Thuringia)
Total (Germany) 0.260 51.39 94.30 81.1 46,414
a
Note: Discounted present value over a five-year period (for people 16 years).
is interesting because it could have been anticipate the positive effects of hosting the
assumed that the effect is similar to Games for the country as a whole.
AFFECT. Sharing results with policy mak- Table 4 includes average and aggregate
ers may be a precondition that results can WTP by postcode area and a description of
affect policy decisions; yet, this is not related the cities and states in those postcode areas.
for the respondents of our survey. The table reveals large differences in average
Several socio-demographic characteristics WTP between regions, ranging from e 31 in
significantly affect the amount of WTP. As Hanover (postcode 3) to e 100 around
in other studies, males report higher WTP Cologne (postcode 5). The significant nega-
than females (Coates and Szymanski, 2015), tive effects in the models for POS_WTP and
age reduces WTP (Walton et al., 2008), and LN(WTP) support the relatively low WTP in
WTP rises with income (Preuss and the Hanover region. Although the Hanover
Werkmann, 2011). People with more educa- region is the third largest region in terms of
tion (SCHOOLING) have higher WTP, per- total population, aggregate WTP is the sec-
haps because they are more likely to ond lowest among all regions.Aggregate WTP is lowest in the Leipzig clubs, e.g., Dortmund, Schalke, Essen,
region (postcode 0). This region also has Duisburg, Mönchengladbach. Residents of
the lowest percentage of respondents (17%) this region still seem to be supportive of the
willing to support the Games. The low sup- Summer Games despite the failed initiative
port may be due to Leipzig’s failed bid for in 2003. Nevertheless, more than two-thirds
the 2012 Summer Games; the city did not of respondents do not support hosting the
even reach the status of an official candi- Games, perhaps because the Ruhr area is
date city. This region is also less affluent included, a working-class region.
than most others and may therefore be less The Cologne region (postcode 5) lies south
willing to pay for large sporting events. For of the Düsseldorf region. Its e 100 average
example, the Leipzig stadium was one of WTP, the highest, coupled with its second-
only two World Cup stadia in Germany for highest population of all German regions
which the federal government had to cover gives it an aggregate WTP of e 10.8 billion,
a large part of the costs. Thus, if Germany by far the highest of all regions. The high
decides to bid for the Summer Games in support might be expected in a region with
the future, the Hanover and Leipzig region several first division teams in football (FC
should not be considered given their rela- Cologne, Bayer Leverkusen), ice hockey
tively low support. (Cologne Sharks, Krefeld Penguins), and
The Leipzig bid for the 2012 Games was handball (VfL Gummersbach). Furthermore,
preceded by an intra-German competition unlike many other major German cities,
with four other potential hosts: Stuttgart Cologne had not recently been disappointed
(postcode 7), Rhine-Main Region (Frankfurt; by failed bids to host the Summer Games.
postcode 6), Düsseldorf Rhine-Ruhr (post- Hamburg (postcode 2) and Berlin (post-
code 4), and Hamburg (postcode 2). The selec- code 1) were considered for the 2024 bidding
tion committee eliminated the first three early process. Public support, in terms of positive
in the process in 2003. Hamburg lost the final WTP and average WTP, is higher in the
round against Leipzig. Average WTP in the Berlin region, despite its failed bid for the
first three regions is about the same, e 35 to 2000 Games. Perhaps the German Olympic
e 39, and low compared with other regions. Sports Confederation should have chosen
Perhaps their earlier, unsuccessful bids con- Berlin as the candidate instead. But, maybe
tributed to weak support in Stuttgart and not; only 28.5% of respondents reported a
Frankfurt. Moreover, Stuttgarters may have positive WTP, far short of the 50% required
been wary of large-scale projects amid long- for a referendum to pass.
lasting controversies surrounding Stuttgart 21, The Munich region (postcode 8) also has
a contentious underground railway and urban a history of recent failed bids, for the 2018
development project. Similarly, ongoing Winter Games and the 2022 Winter Games,
debate about an expansion of Frankfurt’s air- though the 1972 Munich Summer Games
port may have dampened support for yet remain the only German-hosted Olympic
another large project. Games since the Second World War. But
The Düsseldorf region has the strongest most residents today do not remember or
support (31%) for hosting the Games, signif- did not experience the 1972 Games and their
icantly different from the Hamburg region in WTP is close to the overall German average.
the positive WTP model. The higher support The WTP aggregated over all regions
for sport in this region compared with other over a five-year period exceeds e 46 billion
German regions may be because it is home (Table 4), far greater than the estimated e 7.4
to several traditional Football Bundesliga billion cost of hosting the 2024 HamburgGames (Hamburg, 2015). The expectation to host the Games, that aggregate German
was that the federal government would cover WTP exceeds estimated costs, but that using
e 6.2 billion; yet, it only acknowledged this taxes as the payment vehicle would doom
figure without making any commitments. any referendum to host the Games. Results
The aggregate WTP for the Hamburg region revealed large differences in WTP across
(postcode 2) was e 3.9 billion, enough to regions, with the Cologne region reporting
cover the difference between the estimated the highest average and total WTP – a find-
cost and the expected federal contribution. ing which should be considered when select-
Since non-response bias could not be fully ing potential hosts in the future. The
addressed in this study, those who do not perception of positive intangible effects of
care about Olympic Games are likely to have the Olympics (happiness derived from sport-
a lower WTP. To the extent that this is true, ing success, athletes as role models) had a
estimated aggregate WTP is biased upward. positive effect on prestige and WTP.
This study has implications for policy Consequential surveys also lead to greater
makers. Since only 26% of respondents WTP for two reasons: more respondents are
expressed a positive WTP, it will be hard for willing to pay positive amounts and the
a referendum to win a majority in the future. magnitude of WTP is higher.
It may be more promising to use payment The limitations of this study represent
vehicles other than tax increases. Another opportunities for future research. CVM
implication would be to highlight the intan- researchers should examine the sensitivity of
gible effects in the public debate rather than WTP to alternative payment vehicles. Tax
any economic impact estimates when dis- increases as a payment vehicle enhance per-
cussing the possible benefits of hosting. ceived consequentiality which increases
Research on mega events such as the WTP. But, dislike of taxes may decrease
Olympic Games has focused on the poten- WTP. It would be interesting to see if the
tial economic and tourism development likelihood of a positive WTP and the
impacts of hosting (Baade et al., 2010; amount of WTP differ when the scenario
Feddersen and Maennig, 2013), and hosting involves payment vehicles other than tax
in the context of inter-urban competition increases. Similarly, the impact of conse-
(Andranovich et al., 2001; Hiller, 2000; quentiality should also be examined in future
Richards and Wilson, 2004; Whitelegg, studies with varying payment vehicles.
2000). The findings here suggest that a
more powerful (and perhaps realistic) argu- Acknowledgements
ment for hosting may be the public goods The authors would like to thank the students
benefits local residents receive, rather than who assisted with the data collection for this proj-
any forecasted economic impacts. ect. These are Magnus Metz, Jonas Resch,
Ioannis Foukis, Friedrich Mahle, Yogev Cohen,
Christian Saur, Stefan Fink, Florian Deitermann,
Conclusion Fabian Schuler, Julian Schöneborn, Annika
Hentschel, Tobias Freudenthal, Dennis
In light of two failed German referenda on
Lichtenwimmer, Frederic Goller, Nicolaus von
hosting Olympic Games, this study used Schultzendorff, Stephanie Fuchs, Philipp Beuers,
CVM to examine German public support Michael Piechulla, Kevin Sonneveld, Max
for hosting Olympic Summer Games. The Eylmanns, Nermin Mujezinovic, Gavin Sexton,
results suggest that about one-quarter of Stefan Ernst, Elisabeth Schädlich, Jan Tulke,
respondents are willing to pay higher taxes Frederik Schnock and Ryan Viehrig.Funding Carson RT (2012) Contingent valuation: A prac-
This research received no specific grant from any tical alternative when prices aren’t available.
funding agency in the public, commercial, or not- Journal of Economic Perspectives 26(4): 27–42.
for-profit sectors. Carson RT and Groves T (2007) Incentive and
informational properties of preference ques-
tions. Environmental and Resource Economics
References 37(1): 181–210.
Amin A (2003) The economic base of contempo- Carson RT, Groves T and List JA (2014) Conse-
rary cities. In: Bridge G and Watson S (eds) A quentiality: A theoretical and experimental
Companion to the City. Malden, MA: Black- exploration of a single binary choice. Journal
well Publishers Ltd, pp. 115–129. of the Association of Environmental and
Andranovich G, Burbank MJ and Heying CH Resource Economists 1(1): 171–207.
(2001) Olympic cities: Lessons learned from Castellanos P, Garcı́a J and Sánchez J (2011) The
mega-event politics. Journal of Urban Affairs willingness to pay to keep a football club in a
23(2): 113–131. city: How important are the methodological
Atkinson G, Mourato S, Szymanski S, et al. issues? Journal of Sports Economics 12(4):
(2008) Are we willing to pay enough to ‘back 464–486.
the bid’? Valuing the intangible impacts of Coates D and Szymanski S (2015) Willingness to
London’s bid to host the 2012 Summer Olym- pay to host the Summer Olympic Games. Work-
pic Games. Urban Studies 45: 419–444. ing Paper, University of Maryland Baltimore
Baade RA, Baumann RW and Matheson VA County, USA.
(2010) Slippery slope? Assessing the economic Cummings RG, Harrison GW and Rutström EE
impact of the 2002 Winter Olympic Games in (1995) Homegrown values and hypothetical
Salt Lake City, Utah. Région et Développe- surveys: Is the dichotomous choice approach
ment 31: 81–91. incentive-compatible? American Economic
Barlow A and Forrest D (2015) Benefits to their Review 85(1): 260–266.
communities from small town professional del Saz-Salazar S and Rausell-Köster P (2008) A
football clubs. National Institute Economic double-hurdle model of urban green areas
Review 232(1): R18–R29. valuation: Dealing with zero responses. Land-
Begg I (1999) Cities and competitiveness. Urban scape and Urban Planning 84(3): 241–251.
Studies 36: 795–809. Dickinson DL and Whitehead JC (2015) Dubious
Breuer C and Hallmann K (2011) Die gesellschaf- and dubiouser: Contingent valuation and the time
tliche Relevanz des Spitzensports in Deutsch- of day. Economic Inquiry 53(2): 1396–1400.
lands. Bonn: Hausdruckerei des Statistischen Dillman DA, Smyth JD and Christian LM (2014)
Bundesamtes. Internet, phone, mail, and mixed-mode surveys:
Burbank MJ, Andranovich GD and Heying CH The tailored design method (4th ed.). Hoboken,
(2001) Olympic Dreams: The Impact of Mega- NJ: John Wiley & Sons.
events on Local Politics. Boulder, CO: Rienner Downward P and Rasciute S (2010) The relative
Publishers. demands for sports and leisure in England.
Cameron TA and Huppert DD (1989) OLS ver- European Sport Management Quarterly 10(2):
sus ML estimation of non-market resource 189–214.
values with payment card interval data. Jour- Downward P, Dawson A and Dejonghe T (2009)
nal of Environmental Economics and Manage- Sport Economics. Theory, Evidence and Policy.
ment 17(3): 230–246. Oxford: Butterworth-Heinemann.
Carkin DM and Tracy PE (2014) Adjusting for Drabowicz T (2014) Gender and digital usage
unit non-response in surveys through weight- inequality among adolescents: A comparative
ing. Crime & Delinquency 61(1): 143–158. study of 39 countries. Computers & Education
Carson RT (2000) Contingent valuation: A user’s 74: 98–111.
guide. Environmental Science and Technology Feddersen A and Maennig W (2013) Employment
34(8): 1413–1418. effects of the Olympic Games in Atlanta 1996reconsidered. International Journal of Sport Holcomb B (1999) Marketing cities for tourism. Finance 8: 95–111. In: Judd DR and Fainstein SS (eds) The Tour- Federal Statistical Office (2015) Regionaldaten- ist City. New Haven, CT: Yale University bank Deutschland. Available at: https://www. Press, pp. 54–70. regionalstatistik.de/genesis/online (accessed 15 Humphreys BR, Johnson BK, Mason DS, et al. April 2015). (2016) Estimating the value of medal success Groothuis PA and Whitehead JC (2009) The pro- in the Olympic Games. Journal of Sports Eco- vision point mechanism and scenario rejection nomics. Epub ahead of print 27 February in contingent valuation. Agricultural & 2016. DOI: 10.1177/1527002515626221. Resource Economics Review 38(2): 271–280. Humphreys BR, Lee YS and Soebbing BP (2010) Groothuis PA, Mohr T, Whitehead JC, et al. Consumer behaviour in lottery: The double (2015) Payment and policy consequentiality in hurdle approach and zeros in gambling survey contingent valuation. Working Paper No. 15– data. International Gambling Studies 10(2): 04, Department of Economics, Appalachian 165–176. State University. Available at: http://econ. Johnson BK (2008) The valuation of nonmarket appstate.edu/RePEc/pdf/wp1504.pdf (accessed benefits in sport. In: Humphreys BR and 3 June 2015). Howard DR (eds) The Business of Sports. Hamburg (2015) Olympische und Paralympische Westport, CT: Praeger, pp. 207–233. Spiele in Hamburg. Finanzreport. [Olympic and Johnson BK and Whitehead JC (2000) Value of Paralympic Games in Hamburg. Financial public goods from sport stadiums: The CVM report]. Available at: http://www.hamburg.de/ approach. Contemporary Economic Policy contentblob/4612940/data/pdf-finanzreport. 18(1): 48–58. pdf (accessed 11 March 2016). Johnson BK, Groothuis PA and Whitehead JC Hansen E, Fonager K, Freund KS, et al. (2014) (2001) Value of public goods generated by a The impact of non-responders on health and major league sports team: The CVM approach. lifestyle outcomes in an intervention study. Journal of Sports Economics 2(1): 6–21. BMC Research Notes 7: 632. Johnson BK, Mondello MJ and Whitehead JC Harter JFR (2015) Is the wigwam worth it? A (2006) Contingent valuation of sports: Tem- contingent valuation method estimate for a poral embedding and ordering effects. Journal small-city sports arena. Contemporary Eco- of Sports Economics 7(3): 267–288. nomic Policy 33(2): 279–284. Johnson BK, Whitehead JC, Mason DS, et al. Heisey K (2009) Estimating the intangible benefits (2012) Willingness to pay for downtown pub- of hosting the 2016 Olympic and Paralympic lic goods generated by large, sports-anchored Games for potential bid cities: Berlin, Chicago, development projects: The CVM approach. and San Francisco. Dissertation, German City, Culture & Society 3: 201–208. Sport University Cologne, Germany. Kavetsos G and Szymanski S (2010) National Hiller HH (2000) Mega-events, urban boosterism well-being and international sports events. and growth strategies: An analysis of the Journal of Economic Psychology 31: 158–171. objectives and legitimations of the Cape Könecke T, Schubert M and Preuß H (2016) Town 2004 Olympic Bid. International Jour- (N)Olympia in Germany? An analysis of the nal of Urban and Regional Research 24(2): referendum against Munich 2022. Sportwis- 449–458. senschaft 46: 15–24. Hiller HH and Wanner RA (2014) The psycho- Landry C and List J (2007) Using ex ante social impact of the Olympics as urban festival: approaches to obtain credible signals for value A leisure perspective. Leisure Studies 34(6): in contingent markets: Evidence from the 672–688. field. American Journal of Agricultural Eco- Höfler M, Pfister H, Lieb R, et al. (2005) The use nomics 89(2): 420–429. of weights to account for non-response and Liebe U (2007) Zahlungsbereitschaft für kollektive drop-out. Social Psychiatry and Psychiatric Umweltgüter [Willingness-to-pay for Collective Epidemiology 40: 291–299. Environmental Goods]. Wiesbaden: VS.
Liebe U, Preisendörfer P and Meyerhoff J (2011) braucht_den_superstar.pdf (accessed 20 April
To pay or not to pay: Competing theories to 2015).
explain individuals’ willingness to pay for pub- Thornley A (2002) Urban regeneration and sports
lic environmental goods. Environment and stadia. European Planning Studies 10: 813–818.
Behavior 31(1): 106–130. van Deursen AJAM, van Dijk JAGM and ten
Loomis J (2011) What’s to know about hypothe- Kloster PM (2015) Increasing inequalities in
tical bias in stated preference valuation stud- what we do online: A longitudinal cross sec-
ies? Journal of Economic Surveys 25(2): tional analysis of Internet activities among the
363–370. Dutch population (2010 to 2013) over gender,
Morgan OA and Whitehead JC (2015) Willing- age, education, and income. Telematics and
ness to pay for soccer player development in Informatics 32: 259–272.
the United States. Journal of Sports Econom- Vossler CA and Evans M (2009) Bridging the gap
ics. Epub ahead of print 16 October 2015. between the field and the lab: Environmental
DOI:10.1177/1527002515611463. goods, policy maker input, and consequential-
Pawlowski T, Downward P and Rasciute S (2014) ity. Journal of Environmental Economics and
Does national pride from international sport- Management 58(3): 338–345.
ing success contribute to well-being? An inter- Vossler CA and Watson SB (2013) Understanding
national investigation. Sport Management the consequences of consequentiality: Testing
Review 17: 121–132. the validity of stated preferences in the field.
Porter PK and Fletcher D (2008) The economic Journal of Economic Behavior and Organization
impact of the Olympic Games: Ex ante predic- 86: 137–147.
tions and ex post reality. Journal of Sport Walton H, Longo A and Dawson P (2008) A con-
Management 22(4): 470–486. tingent valuation of the 2012 London Olympic
Preuss H and Werkmann K (2011) Erlebniswert Games: A regional perspective. Journal of
Olympischer Winterspiele in München 2018. Sports Economics 9: 304–317.
Sport und Gesellschaft 8(2): 97–123. Whitelegg D (2000) Going for gold: Atlanta’s bid
Richards G and Wilson J (2004) The impact of for fame. International Journal of Urban and
cultural events on city image: Rotterdam, cul- Regional Research 24(4): 801–817.
tural capital of Europe 2001. Urban Studies Wicker P, Kiefer S and Dilger A (2015) The value
41(10): 1931–1951. of sporting success to Germans: Comparing
Rowe D and McGuirk P (1999) Drunk for three the 2012 UEFA Championships with the 2012
weeks – Sporting success and city image. Inter- Olympics. Journal of Business Economics 85(8):
national Review for the Sociology of Sport 897–919.
34(2): 125–141. Wicker P, Whitehead JC, Johnson BK, et al.
Santo C (2008) Beyond the economic catalyst (2016) Willingness-to-pay for sporting success
debate: Can consumption benefits justify a of Football Bundesliga teams. Contemporary
municipal stadium investment? Journal of Economic Policy 34(3): 446–462.
Urban Affairs 29(5): 455–479. Wiser RH (2007) Using contingent valuation to
Schmidt SL and Högele D (2011) Deutschland explore willingness to pay for renewable
braucht den Superstar. Die gesellschafliche
energy: A comparison of collective and volun-
Bedeutung von Vorbildern im Profifußball.
tary payment vehicles. Ecological Economics
Available at: https://www.bundesliga.de/media/
62(3–4): 419–432.
native/dokument/isbs_issue_4_deutschland_You can also read
